2024 ValAct Meeting Virtual - Session 9B: Building AI Reference Tools for Valuation Actuaries - for Safety and Efficiency 9.19.24 Recording

In this session, we will explore the opportunities and challenges of leveraging AI tools to support valuation actuaries in their work. We will begin by introducing the concept of large language models (LLMs), discussing their capabilities, limitations, and potential errors. We will then delve into case studies that demonstrate how LLMs can be effectively utilized to reference actuarial documents. The session will also address key considerations when developing AI tools to mitigate errors, enhance safety and security, and improve overall efficiency. The case studies will cover fine-tuning and information retrieval techniques such as retrieval augmented generation (RAG). Our goal is to provide attendees with a foundational understanding of the nature of LLMs and offer practical insights and ideas for actuaries to build their own processes and tools that ensure accountability, safety, and efficiency.